[jira] [Created] (MESOS-1738) SubprocessTest.Status sometimes terminated by SIGTERM

["Jie Yu (JIRA)" <ji...@apache.org>]

Jie Yu created MESOS-1738:
-----------------------------

             Summary: SubprocessTest.Status sometimes terminated by SIGTERM
                 Key: MESOS-1738
                 URL: https://issues.apache.org/jira/browse/MESOS-1738
             Project: Mesos
          Issue Type: Bug
          Components: libprocess
    Affects Versions: 0.20.0
         Environment: ./configure

clang version 3.3 (tags/RELEASE_33/final)
Target: x86_64-apple-darwin12.5.0
Thread model: posix

Mac OS X 10.8.5
            Reporter: Jie Yu


Sometimes, you'll get this (once every 100 runs):

{noformat}
[==========] Running 1 test from 1 test case.
[----------] Global test environment set-up.
[----------] 1 test from SubprocessTest
[ RUN      ] SubprocessTest.Status
W0827 11:16:26.919852 1972556160 glog.hpp:41] RAW: Received signal SIGTERM; exiting.
W0827 11:16:27.055503 1972556160 glog.hpp:41] RAW: Received signal SIGTERM; exiting.
Terminated: 15
{noformat}

After adding a few prints like the following:

{noformat}
TEST_F(SubprocessTest, Status) 
{
  ...
  // SIGTERM.
  s = subprocess("sleep 60");

  ASSERT_SOME(s);

  LOG(INFO) << ::getpid();
  LOG(INFO) << s.get().pid();
  kill(s.get().pid(), SIGTERM);
  ...
}

inline void handler(int signal)
{
  if (signal == SIGTERM) {
    RAW_LOG(WARNING, "Received signal SIGTERM; exiting. %d", ::getpid());
    ...
  }
  ...
}
{noformat}

I got this:

{noformat}
[==========] Running 1 test from 1 test case.
[----------] Global test environment set-up.
[----------] 1 test from SubprocessTest
[ RUN      ] SubprocessTest.Status
I0827 11:16:27.055408 1972556160 subprocess_tests.cpp:80] 11072
I0827 11:16:27.055503 1972556160 subprocess_tests.cpp:81] 11347
W0827 11:16:26.919852 1972556160 glog.hpp:41] RAW: Received signal SIGTERM; exiting. 11072
W0827 11:16:27.055503 1972556160 glog.hpp:41] RAW: Received signal SIGTERM; exiting. 11072
Terminated: 15
{noformat}

The behavior is similar to this thread:
http://compgroups.net/comp.unix.programmer/sigterm-to-child-also-kills-parent/535939



--
This message was sent by Atlassian JIRA
(v6.2#6252)